Approaches to risk stratification in high risk myeloma

Share :
Published: 31 Oct 2024
Views: 24
Rating:
Save
Dr Irene Ghobrial - Dana-Farber Cancer Institute, Boston, USA

Dr Irene Ghobrial speaks to ecancer about approaches to risk stratification in high-risk myeloma.

Decisions on treating smouldering myeloma focus on whether to start early treatment.

Patients are concerned about progression timelines, highlighting the need for better risk stratification.

With a 10% annual progression rate, current clinical markers help predict outcomes, but dynamic models like PANGEA and genomic analyses provide deeper insights.

Innovations in proteomics and AI aim to enhance risk assessment and treatment strategies for improved patient outcomes.

We know that when we see a patient with smouldering myeloma we have a big decision to make – should we try and treat them early or not? The patients, when they come, they want to understand whether they will progress in the next two years or five years or they may not progress for a very long time. That’s critical for us to have an improved risk stratification for our patients.

In general, if you think about smouldering myeloma it has a 10% chance of progression per year but that’s a heterogeneous population. Some of them will progress very fast, what we call high-risk smouldering myeloma, and some may not progress that fast and they are more MGUS-like. Now, currently we have clinical markers that can help us predict who will progress and we call that the 2/20/20 model, which is the 2g protein, 20% plasma cells in the bone marrow and 20 light-chain ratio. But that’s a static model which means that we don’t know if that person is going to progress rapidly or not. So there are multiple things that can improve on the risk stratification of our patients and this is critical for us to bring those to the patients.

The first one is can we use dynamic models. We brought up the PANGEA model because now you can look at the changes over time of your haemoglobin, of your M spike and light-chain and put that in a calculator and then you can give to that person their own personalised risk of progression. Then you can look at genomics. We use FISH now in the bone marrow biopsy but FISH is not accurate, it’s a 50-year-old technology and the bone marrow biopsies are hard to do and many of our patients don’t want to have it. So can we use circulating tumour cells? Can we use whole genome sequencing to be more accurate in our prediction? In fact, circulating tumour cells are also predictive – if they are high in the blood we know that that patient is likely going to progress.

Then, of course, we look at now we developed something called SWIFT-Seq which is using single cell RNA sequencing instead of DNA whole genome sequencing to try and look at the genomic changes at the RNA level. This is well known, in the old days we used to have the gene expression profiling, now can we bring gene expression profiling along with cytogenetics, along with the circulation or proliferation potential of those cells, to give us some more insight into the risk of progression? All of those things can be done in the clinic.

Then if you stop focussing on the tumour cells and look at the immune cells, you can start asking questions if the immune system is aged or is abnormal or is exhausted will that help me predict this patient will progress or not. Can I use an immune profile that can help me understand whether that patient is going to progress but also can I use it for therapeutics – is that patient going to go on an immunotherapy like a CAR T or a bispecific or not?

Then, finally, things that we need to understand more is proteomics of inflammation, of the microbiome, other things that can be put together to help us understand whether a person is going, or likely going, to progress in the next two years or five years or not.

How do you foresee the landscape changing in the next 5 years?

If I can have the prediction, the best prediction, I think, one, we will have improved risk stratification. We’re doing deep learning, we’re doing AI in everything now. Can we have a better way to look at bone marrow biopsies or circulating tumour cells? Can we have the next generation sequencing in the clinic right now? We should stop using FISH and older technologies and really getting to the next generation. Then can we improve on therapy for our patients? We are starting now to look at bispecifics and CAR T, we think that that’s the future where we use our best drugs, our immunotherapy, as early as possible when your T-cells are not exhausted, when you have less cancer cells.

I really think that treating multiple myeloma after it has already gone into end organ damage is way too late, just like you’re treating with a static cancer. That’s not how we’re going to cure our patients. For any other cancer you never wait for the patients to fall apart and have kidney failure and fractures and then you treat them; you treat it when it’s a truly malignant condition. This is what we’re struggling with, we’re trying to define what is the malignant condition of smouldering myeloma and that’s when we should be treating our patients to give them the best outcome and potentially cure.

So, do I hope that some patients will be MRD negative and cured and potentially without evidence of disease, without therapy, for a long time? My hope is, yes, in five years we will have that.